pic: Blue Lightning 578 presents KateGleasonOne

[KenWittlief]

the ADXRS150EB (eval board version)

when you pull the joystick trigger and close the loop it goes where you point it like the whole bot is one of the hitec servos - its SOOOO nice to drive that way

even with our little 1 inch castors for front wheels :^)

[Scooter]

Originally Posted by KenWittlief

the ADXRS150EB (eval board version)

when you pull the joystick trigger and close the loop it goes where you point it like the whole bot is one of the hitec servos - its SOOOO nice to drive that way

Yep...good times...I've had to use those on several occasions (I am working on an inertial autopilot unit for RC helis right now...have 3 sitting on my desk)....

What do you mean drives like a hitec servo? How so? Does the stick just give the controller a position command? Do you only use it for rotation, or do you use it for linear motion, too?

-Bill

[KenWittlief]

we closed the loop on steering last year by looking at the joystick1 X axis as a desired turn rate, and compairing it to the actual turn rate from the gyro sensor

it worked very well, esp keeping the bot straight going up the ramp- but the thing that didnt sit right with me was, when the commanded turn rate was equal to the actual turn rate, you ended up with no error signal - so there was no obvious way to command the motors to keep turning (if you were turning)

this year for auton mode, we create a compass heading with the yaw sensor by initializeing a variable to 32,000 - then we add the (sensor>>2 -127) to that on every loop of the code - so when the bot turns right, the accumulator goes down, when it turns left it goes up - and we use that as the relative actual heading of the robot

this let us use a straight PID loop to steer - if we want to be going straight we take the error from 32,000 and scale the error signal to be the X axis command into the one joystick equation

then it hit me one day. In normal mode, you can create another variable, desired heading, and init it to 32,000 as well - and when the driver moves the joystick add the x axis signal to the accumulating desired heading - and it will generate a desired (sorta like a projected heading) and we could compair the compass heading, and have the same kind of PID equations we used in auton mode

so in auton mode, the SW has preset desired headings, like go straight (32000) then turn left 80° (new desired heading is 29500) ect

and when the driver closes the loop it starts at 32000, clears the compass to 32000, and as he moves the stick left and right its changing the desired heading in realtime

all you have to do is get your gains right for your drivetrain - I think we are dividing the error signal by 8 - and limit the commands to the 8 bit range - and it works so nice

what I mean by steering like a hitec servo is, its like you connected a servo to the x axis of the joystick, thats how the robot turns - you nudge the stick slightly to the left and that desired heading starts accumulating to the left, and the robot slowly and precisely turns to the left

you hold the stick straight and push it forwards and the bot move forwards, dead straight.

[ngreen]

If if worked fast enough that could be a quick 80 pts. Watching practice round the bonus balls didn't get touched in 95% of the matches.

I want the job to reset the bonus ball.

[KenWittlief]

I dont think we would take the time to try to acquire other 10 point balls off the tees - think about it, it would take several seconds each to aquire them, and drop them into our hopper

in that amount of time we could get 5 or 6 5 point balls from the floor

from watching the games so far, it looks like being able to focus on the biggest bang for the buck, as it were, is whats going to win the game - I think for us that will be sweeping the 24 balls off the floor and getting them to the HPs - if there are 10 pointers scattered in there, thats great, but we will not go out of our flightpath to chase them.

[ngreen]

Yeah I wouldn't either. In my robot's case where we can only move basically one at a time I was looking at ways to maximize that potential, ie if you can move one 10 pt faster than 2 5 pts move the 10 pt. We can't do it. I didn't realize the bar was that close to the top of the bonus balls.

I've got better strategy ideas than that though. I keep those secret because they are more robot specific. They're easily countered enough so they can be used for a couple matches.